Elderly adults with diabetes are disproportionately burdened with dementia. They also experience reduced longevity as compared with non-diabetics, with a particularly increased mortality recently noted among those who were normal-weight prior to the development of diabetes. It remains unclear, however, what factors contribute to these differences. Some studies suggest that dementia may represent another manifestation of small-vessel disease in diabetics, often co-existing with other microvascular complications, and resulting largely from the development of microvascular lesions in the brain (micro- and lacunar infarcts), and less commonly from the lesions of Alzheimer's disease. Adults who develop diabetes despite a history of normal weight tend to have central obesity, with increased visceral adiposity, which also appears to increase the risk of microvascular complications systemically. The goal of the proposed research is to investigate the associations of obesity and diabetes with cognition in the elderly, longevity, and the neuropathologic correlates of dementia at autopsy. A specific focus of our study will be to understand how body composition during the life course may influence cognition and longevity in different subtypes of diabetes. This project will utilize existing data from the Honolulu-Asia Aging Study (HAAS) - a prospective cohort of 3734 men followed for more than 20 years, with available brain autopsy information on 852 men. We hypothesize that normal-weight diabetics carry an excess burden of small-vessel disease systemically, which will manifest as an increased risk of cognitive decline and dementia, reduced longevity, and an increased frequency of cerebral microvascular disease and associated atrophy at autopsy, as compared with overweight/obese diabetics. These efforts are expected to: (1) clarify mortality differences among diabetes subtypes defined by weight history and body composition, (2) determine whether risks of dementia and cognitive impairment and decline are increased among normal-weight diabetics, and (3) identify differences in the burden of dementia-related brain lesions for autopsied decedents according to diabetes and weight history during life. An understanding of the processes underlying cognitive and longevity differences among diabetics is crucial to identifying targets for disease prevention and risk modification. Our hypotheses, compatible with information from limited previous studies, present a novel framework for understanding how the underlying pathogenic vascular processes common to diabetes may vary by weight status.